Re: [rfc] [2/4] SPU overlay support: The SPU target part

["Ulrich Weigand" <uweigand at de dot ibm dot com>]

Daniel Jacobowitz wrote:
> On Fri, May 11, 2007 at 12:20:05AM +0200, Ulrich Weigand wrote:
> > This makes it possible to use the remaining slots to hold additional
> > information needed to handle return jumps crossing an overlay 
> > boundary.  In those cases, the slots are set up to hold:
> >   [0] Return stub entry point in the overlay manager
> >   [1] Partition number of the overlay section to be returned to
> >   [2] Actual return address in the (restored) overlay section
> 
> Clever.  Could we have a comment about this in GDB somewhere?
> Apologies if there was one; I didn't see it.

Sure, I'll add a comment.

> > > This is clever, but kind of sneaky.  We show signal return trampolines
> > > and dummy call trampolines, so I'm not sure why it's necessary to
> > > hide overlay return stubs.  Do you think this is more useful than
> > > confusing?
> > 
> > Both signal return and dummy call trampolines are entities the
> > user actually knows about and wants to see.  The overlay mechanism
> > is supposed to be fully transparent to the user; I'd compare the
> > overlay call and return stubs to things like PLT stubs in ELF
> > -- we don't show those either.
> 
> Well, they never end up on the stack.  We would if they did.  But this
> isn't a big issue to me; I'd be very confused if I stepi'd a
> return instruction and ended up somewhere other than the function
> listed in the backtrace, but my use of GDB is probably not typical.

The return stub isn't on the stack either.  We'd have to manufacture
an extra zero-sized stack frame between caller and callee, with 
special unwind rules to be able to continue the backtrace.  I guess
this would be possible somehow, but I still don't see that this would
provide any benefit to most users ...

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU Toolchain for Linux on System z and Cell BE
  Ulrich.Weigand@de.ibm.com